Jökull - 01.12.1990, Blaðsíða 106

in Figs. lb to lf. The greatest uncertainty is probably
in the prediction of the excess S042- and Ca2+ as
discussed above.
METHODS
Snow cores were collected from the westem part
of the Vatnajökull glacier during the spring expedition
of the Icelandic Glaciological Society, 1988. The drill
sites were located on the glacier by the Loran-C loca-
tion system. The Loran-C coordinates are converted to
conformal conic Lambert coordinates. The locations
close to the Grímsvötn area (cores 1, 2 and 3, Table I)
are corrected by moving them 340 m to the west but
others are moved 340 m to the west and 80 m to the
north (Fig. 2 and Tables I, II and III). This is done to
compensate for the local effect of the Grímsvötn area
on the Loran-C signals (Fig. 2). Experience has shown
this to be the appropriate correction (Finnur Pálsson
and Helgi Björnsson, 1990, pers.comm.). The eleva-
tion of the drill holes is calculated from their location
using the digitizedmaps of Bjömsson (1988) for cores
4 to 9 (Fig. 2, samples 4.1 to 9.1 in Table I) and un-
published digitized maps (Helgi Bjömsson and Finnur
Pálsson, 1990, pers. comm.) forcores 1 and 3 (Fig. 2,
sample 1.1,1.2,1.1,1.2,1.3.. etc, 3.1 and3.2, Tables I
and II). The elevation of core 2 (Fig. 2, sample 2.0 and
2.1,2.2.. etc. inTable I andlll) is obtained by a level-
ling measurement to a permanent bench-mark on the
Grímsfjall mountain (Fig. 2, Theódór Theódórsson,
1990, pers. comm.).
The snow cores were collected by PICO
lightweight hand coring auger. The core barrel, ex-
tensions and T-handle are made of fiber glass but fit-
tings, adapters and cutting head are machined from
aluminum (Koci, 1983). Drilling through the 1987-
1988 precipitation layer was carried out in several
steps. The auger cuts through 50 cm up to a me-
ter each time, it is then extracted from the hole and
the core and snow-cuttings placed on a core-shaped
polyethylene platter (Fig. 3). The snow-cuttings are
discharged, the length of the core measured and the
integrated core length compared with the down hole
logging. The snow in the first meter of the core is
compressed during coring, causing the measured core
length to be shorter than the hole depth. This discrep-
ancy is corrected for by lowering the integrated core
by the amount of the snow compression in the first
meter (Figs. 4, 5, and 6). Thus, the cores are shown to
start at 10 to 30 cm depth in the drillholes in Figs. 4,
5, and 6, but it is important to note that the cores are
taken from the surface and down (Tables I, II and III).
Each 25 cm part of the core is split into two. One half
is placed in a 1 liter polyethylene sample container
but the other, which at the end of coring represents
the integrated core, is placed in a 6 liter polyethylene
bucket. The 25 cm sample is taken for studying down-
hole core changes but the integrated sample to study
spatial changes in the snow chemistry on the glacier.
From cores 3 to 9 only integrated samples were taken
(Table I).
After the expedition, the samples were melted in
the laboratory. This occurs upon storage for up to a
week at room temperature. After melting, about 50 ml
of each sample were used to obtain the pH of the melt-
water, but the rest filtered through a 0.1 prn pore size
filter and placed in prewashed low density polyethy-
lene plastic bottles. The analytical methods and the
average relative analytical errors obtained from dupli-
cate determinations are shown in Table IV. The rela-
tive error shown in Table IV for Cl“ and S042- is the
maximum difference between two of up to 16 repli-
cates. The relative error is high reflecting the low
concentration of the ions and that the measurements
are done close to the detection limits formost methods.
The eluant for the determination of Cl_ and S042- is
potassiumhydrogenphtalate (1.25 g/1). The pH of the
eluant is adjusted to 4.33 by 0.1 M potassium hydrox-
ide. Before sample injection, a concentrated eluant
is added to the samples in order to compensate for
sample effect on the spectra. A relatively large sam-
ple loop (500 //1) is used to enhance the sensitivity
of the method. The eluant used for the measurement
of sodium and potassium ions is deionized water with
pH adjusted to 4.5 by purified nitric acid and samples
are treated as described above for chloride and sulfate.
The analytical results are tested using charge balance
equations (1) and (2).
The concentration of bicarbonate is calculated at
22 °C from the measured pH assuming the partial
pressure of CO2 to be 10-3 5 bar using the equilibrium
102 JÖKULL, No. 40, 1990